Global Equity Allocation

(1)

msci.com

Index Report

Analysis of Issues Related to Geographic Allocation of Equities

Prepared for the Ministry of Finance of Norway

(2)

Contents

Contents ... 2

Executive Summary ... 4

Section I: The Globalization of Equity Portfolios ... 6

A Globalized Opportunity Set: Reducing Home Bias ... 6

The Expansion of the Global Opportunity Set: Evolution of Market Portfolio ... 7

Globalization and Liberalization ... 8

The Increasing Role of Emerging Markets ... 10

Changing Approaches to Equity Allocation ... 12

Summary ... 15

Section II: Alternative Weighting Schemes: Market-Cap

Weighting versus GDP Weighting ... 16

Market Capitalization Weighting versus Alternative Weighting Schemes ... 16

GDP Weighting of Global Equities ... 16

Risk and Return ... 19

Attributing Returns ... 22

A Discussion of GDP Weighting ... 25

Summary ... 26

Section III: An Examination of Concentration Risk with a Focus

on the US and Europe ... 27

The Issue of Geographical Concentration Risk ... 27

Return and Risk of US and European Equities ... 28

Simulating Over- and Under-weights to the US and Europe ... 30

Attributing Return and Risk ... 32

Foreign Exposure of US and European Equities ... 36

Valuations ... 37

Summary ... 39

Section IV: Emerging Markets ... 40

The Rise of Emerging Markets Since the 1980s ... 40

(3)

Return and Risk of Emerging Markets ... 46

Fundamental Characteristics of Emerging Markets ... 50

Simulation Results: The Impact of Allocating More Towards Emerging Markets100 Institutional Trends in Emerging Market Allocations ... 54

Summary ... 56

Author Information ... 57

References ... 58

Appendix A: Currency Impact on International Investments 60

Appendix B: The Barra Global Equity Model ... 62

Introduction to Equity Factor Models and the Barra Global Equity Model ... 62

Forecasting Global Equity Portfolio Risk ... 64

Estimation Universe ... 66

GEM2 Factor Structure ... 67

Estimation of GEM2 Factor Returns ... 74

Formation of the GEM2 Factor Covariance Matrix ... 76

The GEM2 Specific Risk Model ... 77

Appendix C: A Discussion of Full Market Cap Weights... 79

Appendix D: Regime Analysis for US and Europe ... 82

Recessions ... 82

Market Volatility ... 83

Asset Price Bubbles ... 85

Appendix F: Stock-Level Concentration Risk ... 88

Appendix G: Additional Detail on Sections III and IV ... 90

Local Currency Results ... 90

Simulating Over- and Under-weights to the US and Europe ... 92

Decomposition of Returns Into Long-Term Components ... 98

Client Service Information is Available 24 Hours a Day ... 102

Notice and Disclaimer ... 102

(4)

Executive Summary

This report discusses the key issues regarding the equity allocation of portfolio capital across different geographical areas. It draws on insights gathered globally by MSCI through interactions with large pension funds and sovereign wealth funds around the world as well as its own research in this area. Focus is paid to analyzing issues from the perspective of a long term institutional investor with sizable assets.

Here, we discuss important criteria and considerations for allocating institutional equity investments across regions/countries. Among the topics highlighted in this report are:

• The implications of weighting schemes including those based on market-capitalization and GDP

• The implications of having greater/lower concentration of the portfolio in US and Europe equity investments

• The characteristics of emerging markets and their risks and rewards

The globalization of economies and the increased integration of capital markets have prompted many institutional investors to rethink their equity investment process. As asset growth is the main objective of equity allocation, biasing it towards the domestic market may come with potentially huge

opportunity costs. Increasingly, institutional investors are shifting from the traditional domestic versus international divide towards an integrated global perspective and a reduction in “home bias”. In this new paradigm, the full equity opportunity set spans the globe. For investors with very long horizons in particular, the opportunity set not only spans the globe but also spans micro caps and frontier markets because with long-term growth as the objective, the natural starting point is an opportunity set that is defined as broadly as possible. Section I of this report provides a comprehensive look at this issue of globalization and the changing landscape of institutional equity investing.

Having established in Section I the rationale for using the global opportunity set as the starting point for asset allocation, Section II examines alternative weighting schemes to determine the geographic asset allocation. While market cap weights are a natural starting point for several reasons, both theoretical and practical, alternative weighting schemes may be attractive for certain reasons. In particular, investors concerned about price inefficiencies or temporary price disequilibria or volatility of security and sector weights may prefer a weighting scheme not based on prices.

Section II focuses specifically on the use of GDP weights to set the country weights in the equity portfolio. Proponents of GDP-weighting highlight several advantages. First, it weights countries based on their economic size, not on the size of their markets, which can sometimes reflect a narrower view of the economy particularly in emerging markets. Second, GDP weighting is not dependent on stock prices which means countries which experience a temporary bubble will not be as heavily weighted as they would be in a market-cap-weighting scheme. Third, GDP weighting has provided increased exposure to emerging markets thus benefiting from the emerging market risk premium. Nevertheless, the success of GDP weighting assumes that economic growth will translate into market returns and that the markets under consideration are sufficiently large, liquid, and representative of their economies. In sum, GDP weights provide another perspective to geographic allocation highlighting the fact that while market cap weights are a natural starting point, an institution’s actual allocation depends on its constraints,

preferences, objectives and investment beliefs on additional sources of return.

Section III explores an important dimension to evaluating different geographical weighting schemes – the concentration risk implied by the actual weights themselves. The US for instance comprises roughly 45% of the world’s market cap currently; Europe on the other hand makes up only 25%. This raises the question of whether using a market-cap weighting scheme exposes the investor to significant

(5)

concentration risk. We delve into the US and Europe cases in greater detail, examining a range of criteria which reflect the risks of these two entities. In our analysis we treat Europe as a single bloc given its integrated economy. While there remains a risk to the continuation of the Euro zone, at least in its current form, for this analysis we assume that the countries that make up Europe will continue to have close economic ties.

Overall, we find that both markets are broad and well-diversified with more similarities than differences. Contagion risks from foreign sources are important for both US and Europe since they have a high exposure to economies abroad. Sovereign event risk, or so-called country factor risk, can also not be diversified away, and to a large extent, cannot be forecast with any great deal of certainty. However, over the long run, both markets have been among the least volatile. Moreover, political and legal institutions are among the most robust in the world, and corporate governance of higher quality. In sum, while there is unavoidable macroeconomic concentration risk in having large weights in the US and Europe, at the same time there are good arguments for having large weights in these markets. Still, some investors may have reason for reducing concentration risk, particularly regarding the current US weight in a market-cap-weighted index. For instance, GDP weighting would result in a lower weight of only 28% to US equities.

Finally, we look at emerging markets equities in Section IV. Emerging markets today constitute a non-negligible part of the opportunity set for global investors. In recent decades, they have allowed investors to take advantage of the relatively greater set of economic growth opportunities in the developing world. Proponents of emerging markets argue for taking advantage of higher growth rates in these markets and a potential risk premium captured by emerging markets. In fact, because forecasts of economic growth do not take into account increases in free float through the effect of market

liberalization on ownership structure, they may underestimate the actual growth potential. Proponents also point out the potential diversification benefits emerging markets may bring.

The main risk to emerging market investing remains the possibility that the previous trend of

globalization may reverse. In addition, emerging markets are more volatile and can suffer long periods of disruption such as the late 1990s during the Russian Ruble and Asian crises. Finally, governance dimensions such as investor protection and shareholder rights are generally less established relative to their developed market counterparts. In sum, emerging markets are a critical part of the global

opportunity set comprising 12% of the global market capitalization currently. We note that if economic weight rather than market weight is used, emerging markets would have a weight of 29%, and using full market cap instead of free float market cap would result in a weight of 16%. Allocations such as these may be considered for investors who are less sensitive to short-term volatility and desire greater exposure to potential long-term economic growth.

(6)

Section I: The Globalization of Equity Portfolios

Section I focuses on understanding why the global equity opportunity set is the natural starting point for the equity allocation of a long term investor.

A Globalized Opportunity Set: Reducing Home Bias

Over the past few decades, globalization of economic activity and increased integration of capital markets have led to a dramatic expansion of the equity universe for international investors. Institutional investors now can access a deeper and broader global equity opportunity set.

The globalization of the equity opportunity set allowed institutional investors to expand their equity investment universe and allocate assets to international equities.1 However, most institutional investors, and European ones in particular, continue to maintain an investment process that separates equity policy portfolios into domestic and international equities at the strategic level, with a significant “home bias” that over weights domestic or European equities.2

Exhibit 1 presents the current levels of home bias in selected European equity markets, as well as the US and Japan, using data from the Coordinated Portfolio Investment Survey (CPIS) conducted by the IMF. The data reveals significant levels of home bias in these major markets, with Japan being the most home-biased, and the US and UK both exhibiting a level of home bias around 52% in 2010. However, Exhibit 1 also shows a decline in home bias over the last decade. In addition, a number of large and leading global pension plans in recent years have moved to a framework where Global Equity is viewed as a single strategic asset class.

Exhibit 1: Equity Home Bias in Selected Countries

Source: IMF (CPIS), MSCI. Home bias is defined as 1 - (actual international equity allocation

/ market-cap based international equity allocation). 2010 numbers are based on the preliminary CPIS data.

1

See Aylur Subramanian, Nielsen, and Fachinotti (2009) 2_{For more discussion on this topic, see Kang and Melas (2010).}

Country 1997 2001 2004 2007 2010* Denmark 79.7% 56.1% 51.4% 48.1% 44.0% Finland 94.1% 74.4% 51.3% 48.9% 40.3% France 83.5% 69.4% 59.5% 65.2% 58.3% Germany NA 49.9% 43.3% 42.6% 42.0% Netherlands 70.2% 35.4% 20.1% 11.1% 2.7% Norway 84.6% 50.4% 46.1% 46.7% 29.6% Sweden 79.2% 51.1% 50.8% 49.8% 49.6% Switzerland NA 57.3% 52.6% 52.1% 51.2% United Kingdom 75.9% 64.0% 56.1% 55.8% 51.6% USA 79.0% 69.6% 59.1% 58.8% 52.8% Japan 92.1% 86.1% 84.7% 85.0% 79.7% Canada 75.4% 59.7% 63.7% 66.1% 68.4%

(7)

The Expansion of the Global Opportunity Set: Evolution of Market Portfolio

The potential benefits of global investing are grounded in modern portfolio theory -- in particular the Capital Asset Pricing Model (CAPM).3 In its original form, the Capital Asset Pricing Model (CAPM) suggested that all investors hold a combination of the risky market portfolio and cash, depending on their risk tolerance.4 Although, the market portfolio - which was defined as a combination of all risky assets imaginable, including equities, fixed income, human capital, etc. - was neither observable nor investable, proxies for the market portfolio were developed.5 The CAPM (which originally covered only domestic assets) was quickly extended to an international version of the framework, the International CAPM (I-CAPM).6 According to the I-CAPM, in an efficient and integrated world capital market, the global market portfolio would replace the domestic market portfolio implying that domestic allocations should not exceed the relative country share in the global market portfolio. Since the mid 1970s, broad global indices like the MSCI World Index (developed markets only) and later the MSCI ACWI (developed and emerging markets beginning in 1987) have been used as proxies for the global market portfolio. Exhibit 2 provides the evolution of global market portfolio as seen by the evolution in the MSCI Global Equity Indices. When MSCI started calculating indices in 1970, the global universe covered only 16 countries consisting of the main developed markets. In the 1980s, with the expansion of global

investing, coverage was expanded to 35 countries and the launch of the MSCI Emerging Markets Index. In the 2000s, as global investors started to explore additional opportunities, MSCI responded by significantly extending its coverage once again and launching the Gulf Cooperation Council (GCC) and the Frontier Markets (FM) indices.

Exhibit 2: The Number of Countries and Securities Included in the Global Equity Opportunity Set*

Source: MSCI, data as of year end. Note: * MSCI EAFE Index (for security data) plus USA and Canada (for country data) in 1969, MSCI World Index between 1974 and 1989, MSCI ACWI Index in 1999, MSCI ACWI + Frontier Markets IMI in 2009, MSCI ACWI + Frontier market All Cap Index in 2010 and 2011.

Coverage has also increased in depth as smaller companies were increasingly considered suitable for international investing. The number of securities covered by the MSCI indices has grown from 600 to more than 14000 in the last forty years. Small cap indices are available for all the countries MSCI covers including emerging and frontier markets and micro cap indices for all developed markets countries.

3_{Markowitz (1952)}

4_{Sharpe (1964) and Linter (1965)} 5

In the U.S. investors used the S&P 500 and later the broader Wilshire 5000 as an investable proxy for their market portfolio. The S&P 500 is a market capitalization weighted portfolio of the U.S.'s largest stocks and covers approximately 75% of the U.S. equity universe. The Wilshire 5000 index is the broadest index for the U.S. equity market, measuring the performance of all U.S. equity securities with readily available data. Dennis A.Tito, A new capital market index (1974)

6_{Adler & Dumas, 1983; Solnik, 1977; Stulz, 1981; Wheatley, 1988.}

1969 1974 1979 1989 1999 2009 2010 2011

Number of countries in the

global opportunity set* 16 19 19 35 47 70 71 70 Number of securities in the

(8)

Globalization and Liberalization

Globalization has been one of the most significant trends of the last 40 years. Over this period, trade barriers and tariffs have been greatly reduced or eliminated altogether, foreign direct investments have increased steadily and the scope of companies operations have expanded well beyond local boundaries. In parallel, capital markets around the world have become more accessible and efficient for foreign and domestic investors alike and a common Financial Accounting Standard has been adopted across more than 100 countries.

Illustrating this trend, the Exhibit 3 shows how world exports as a percentage of global GDP have nearly doubled in the last 20 years. One of the main factors behind this increase in exports has been the gradual removal of trade barriers, as illustrated by the decline in average applied tariffs from 26.3% in 1986 to 8.6% in 2009.

Exhibit 3: Share of Exports in the World’s GDP and Average Applied Tariffs, 1960 – 2010

Source: World Bank, UNCTAD

As the importance of foreign business grew, many companies moved from being exporters to set up full-scale operations that take full advantage of opened economies. Others shifted their production sites to take advantage of lower costs, or sought access to supplies of natural resources.

This move by companies outside of their home country is illustrated by the trend in FDI (Foreign Direct Investment). As seen in Exhibit 4, incoming FDI in the world, particularly emerging economies, has been growing over the last decades, albeit subject to global investment cycles.

0 5 10 15 20 25 30 0 5 10 15 20 25 30 35 1 9 6 0 1 9 6 2 1 9 6 4 1 9 6 6 1 9 6 8 1 9 7 0 1 9 7 2 1 9 7 4 1 9 7 6 1 9 7 8 1 9 8 0 1 9 8 2 1 9 8 4 1 9 8 6 1 9 8 8 1 9 9 0 1 9 9 2 1 9 9 4 1 9 9 6 1 9 9 8 2 0 0 0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 A v e ra g e A p p li e d T a rr if fs ( % ) E x p o rt s a s % o f G D P Exports as Percentage of GDP

Average Applied Tarrif f (Right Axis)

(9)

Exhibit 4: Foreign Direct Investment: Total Flows and Flows to Developing Economies

Source: UNCTAD

As a result of these fundamental transformations in the world economy and in the way companies operate, it has become increasingly difficult to disentangle companies from their global footprint. Exhibit 5 shows the percentage of foreign sales against total sales, as well as the percentage of foreign assets compared to countries in the MSCI World Index. Investors in these countries can take on significant international exposure indirectly even with purely domestic allocations.

1 10 100 1,000 70 72 73 74 75 76 77 79 80 81 82 83 84 86 87 88 89 90 91 93 94 95 96 97 98 00 01 02 03 04 05 07 08 09 10 U S D o ll a rs a t c u rr e n t p ri c e s i n b ill io n s

FDI Flow into Developing and Transition Economies

(10)

Exhibit 5: Foreign Sales and Assets as a Percent of Total Sales and Assets for MSCI World Countries

Source: MSCI, Worldscope. Data as of 2011

The Increasing Role of Emerging Markets

The increased integration of economies and markets globally has resulted in a shift in the balance of economic weights from the traditional developed economies to developing countries. As a group, the latter have gained weight both in market capitalization and in contribution to the world economy. In only 20 years as shown in the graph below, the real GDP share for emerging countries has more than doubled from 12% in 1969 to over 25% today.

Country Foreign Sales as

% of Total Sales Foreign Assets as % of Total Assets Austria 34.7% 24.0% Belgium 59.1% 60.2% Sw itzerland 64.1% 52.9% Germany 48.0% 36.0% Denmark 56.3% 37.9% Spain 39.8% 17.1% Finland 42.8% 30.4% France 45.3% 24.2% United Kingdom 53.9% 45.3% Greece 50.4% 42.0% Ireland 68.2% 56.0% Italy 27.9% 19.8% Netherlands 57.7% 35.1% Norw ay 30.3% 36.5% Portugal 24.0% 17.0% Israel 55.6% 41.9% Sw eden 43.5% 28.9% Australia 31.8% 22.7% Hong Kong 26.3% 28.2% Japan 32.7% 18.4% New Zealand 6.0% 4.0% Singapore 32.7% 22.5% Canada 45.3% 39.6% United States 39.5% 28.9%

(11)

Exhibit 6: Share of real GDP of emerging and developed markets, history and projections

Source: World Bank, USDA

Exhibit 7 details the relative contribution to the World GDP of developed and developing economies by focusing on the 10 largest economies. Based on USDA projections, 20 years from now all four BRIC (Brazil, Russia, India and China) countries will be in the top 10 economies as measured by their nominal GDP, highlighting for developed market investors that “growth may be elsewhere”.

Exhibit 7: Top Ten GDP Weights: Past, Present and Future?

Source: World Bank, USDA. Note: *Data for 2011 is as of May 31, 2011 ** Projected

Developed Emerging 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0% G D P ( W t% )

Country GDP Wt Country GDP Wt Country GDP Wt

1 United States 30.1% USA 25.4% United States 22.2%

2 Japan 16.2% China 10.2% China 16.9%

3 Germany 6.6% Japan 9.6% India 6.1%

4 United Kingdom 4.9% Germany 5.8% Japan 5.5%

5 France 4.5% France 4.5% Germany 4.2%

6 Italy 3.9% United Kingdom 3.9% United Kingdom 3.7%

7 Canada 2.3% Brazil 3.6% Brazil 2.7%

8 Brazil 2.1% Italy 3.6% Italy 2.2%

9 Spain 1.8% India 2.9% Russia 2.0%

10 Russia 1.7% Canada 2.7% Canada 2.0%

2011* 2030**

Rank

(12)

Changing Approaches to Equity Allocation

In the context of a multi-asset class portfolio, the policy objective of the equity allocation is generally asset growth maximization. Over the last 40 years, and in spite of two major market crises in the last 10 years, the cumulative return of equities has been higher than for a global bond portfolio, although with higher volatility (Exhibit 8).

Exhibit 8: Cumulative Returns of Developed Global Equities and Bonds (1969 to 2011*)

Source: MSCI. The cumulative returns for developed market bonds is constructed using the long term government bond yields for 10 countries from the OECD. Returns are in USD. *Data for the bond series ends in 2010 in this exhibit.

In the US, institutional investors traditionally allocated the majority of equity investments domestically. There were several rationale for this. First, US equity markets dominated the global equity opportunity set; its share of the MSCI World Index was 70% in 1970. Second, international equity markets were viewed as difficult to access. Third, domestic equities were seen as a better match for domestic

liabilities. Finally, currency risk was considered an unavoidable part of international equity investing. In the last few decades, these rationales have become harder to sustain. US markets have fallen in their share of global market cap, accessibility to international capital markets has generally improved7, domestic equities have demonstrated that they provide no better link to domestic pension liabilities than global equities, and short term currency risk has been made manageable through the availability of hedging instruments8. For more detailed discussion on long term currency risk, please refer to Appendix A.

7_{Barriers to foreign investment have been lifted or reduced in most countries and market infrastructure improvements have contributed to lower costs and lower} operational risks. Most developed markets now trade at similar levels of efficiency. The consolidation of stock exchanges (for example NYSE Euronext or Nasdaq OMX) and competition within markets are likely to accelerate this trend.

8

The long-term hedged and un-hedged non-U.S. equity performance for U.S. investors has been quite similar, validating the argument that prices of cross-border real assets tend to equilibrate over time. For investors with long horizons, currency risk has been to shown to be relatively less important than equity risk.

50 500 5000 1 9 6 9 1 9 7 0 1 9 7 1 1 9 7 2 1 9 7 3 1 9 7 4 1 9 7 5 1 9 7 6 1 9 7 7 1 9 7 8 1 9 7 9 1 9 8 0 1 9 8 1 1 9 8 2 1 9 8 3 1 9 8 4 1 9 8 5 1 9 8 6 1 9 8 7 1 9 8 8 1 9 8 9 1 9 9 0 1 9 9 1 1 9 9 2 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 8 1 9 9 9 2 0 0 0 2 0 0 1 2 0 0 2 2 0 0 3 2 0 0 4 2 0 0 5 2 0 0 6 2 0 0 7 2 0 0 8 2 0 0 9 2 0 1 0 2 0 1 1

Cumulative Returns f or Developed Markets Bonds MSCI World Index (Gross Returns)

(13)

With the exception of the first, similar changes have also occurred for European institutional investors. In Europe, historically there has been a regional approach to equity allocation, which uses preset fixed weights for the various regions , and which now faces fundamental challenges. This regional approach may reflect, in part, the historical path that European institutional investors took in expanding their equity opportunity set, first from domestic equities to European equities, then to international developed markets with regional mandates, and finally to emerging markets. It may also reflect the perception that regional/country factors are of foremost importance to equity allocation, which was historically the conventional wisdom.

In today’s more global economy, there is growing evidence that regional equity markets are converging, and that country factors are becoming relatively less important in driving the variations in global security returns. Exhibit 9 shows the proportion of return cross-sectionally that is explained by countries,

industries, and styles (risk premia).9 Over the last decade, global industry and style factors have been more important in certain years than country factors in explaining developed market returns. Meanwhile, in emerging markets, the story is quite different; here, country factors have been and remain more important than industry and style factors.

Exhibit 9: Risk and Return Drivers in Developed and Emerging Markets

Source: MSCI

The increasing importance of industries as determinants of global security returns can also be seen in the heterogeneity of sector returns in the last decade. Exhibit 10 shows the dramatically different performance of the three best-performing global sectors (Energy, Materials, and Consumer Staples) and the three worst-performing sectors (Information Technology, Telecommunications Services, and

Financials) over the last ten years, relative to global equities (as measured by MSCI ACWI).

9

Countries, industries, and styles (or risk premia) are identified using the Barra Global Equity Model (GEM2). There are 134 factors in the model. Here, we take the standard deviation of cross-sectional component returns for individual stocks, where the component returns are the contribution to return to each stock from each source of risk. The ratios of the computed standard deviation to the whole are shown. Standard deviations are computed with a weighting scheme based on market capitalization of the individual stocks. The final measure shown in the exhibit is the contribution from factors to cross-sectional volatility (CSV), also called cross-sectional standard deviation or cross-sectional dispersion. Note that the ratio is smoothed by taking the average over the past 12 months. .

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Ja n -9 7 Ja n -9 8 Ja n -9 9 Ja n -0 0 Ja n -0 1 Ja n -0 2 Ja n -0 3 Ja n -0 4 Ja n -0 5 Ja n -0 6 Ja n -0 7 Ja n -0 8 Ja n -0 9 Ja n -1 0 Ja n -1 1 Ja n -1 2

Contribution of Risk Factors to Explained Cross Sectional Volatility (CSV) Emerrging Markets

Styles Countries Industries 0% 10% 20% 30% 40% 50% 60% 70% 80% Ja n -9 7 Ja n -9 8 Ja n -9 9 Ja n -0 0 Ja n -0 1 Ja n -0 2 Ja n -0 3 Ja n -0 4 Ja n -0 5 Ja n -0 6 Ja n -0 7 Ja n -0 8 Ja n -0 9 Ja n -1 0 Ja n -1 1 Ja n -1 2

Contribution of Risk Factors to Explained Cross Sectional Volatility (CSV) Developed Markets

(14)

Exhibit 10: Performance of Selected Global Sectors Relative to Global Equity

Source: MSCI, performance is in USD.

With industry membership becoming more important to stock returns, country membership has necessarily become less important. Not surprisingly, correlations between European equities and other regional equity markets have risen significantly over the last decade (Exhibit 11).

Exhibit 11: Rising Correlation between European Equities and Other Regional Equities

Source: MSCI 0 50 100 150 200 250 300 350 98 99 00 01 02 03 04 05 06 07 08 09 10

Energy Materials Consumer Staples

Information Technology Telecommunication Services Financials

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 74 77 80 83 86 89 92 95 98 01 04 07 10

(15)

The increasingly global nature of business activities and the increased integration between global financial markets has led to a marked convergence among regional equity markets, even if the recent global crisis has exacerbated the extent of the convergence. This trend has weakened the traditional framework of regional allocations used by European investors. Moreover, as asset growth is the main objective of equity allocation, biasing it towards the domestic market may come with potentially huge opportunity costs.

In this new paradigm, the full equity opportunity set spans the globe and is the natural starting point for equity allocation. A representative benchmark or index must be similarly broad but still investable. The MSCI Global Investable Market Indices Methodology was developed with this goal in mind. Today, the MSCI ACWI spans both developed and emerging markets and encompasses 45 countries.

Summary

The last few decades have seen the increasing globalization of economies accompanied by the integration of financial markets. This trend has led to a marked convergence among regional equity markets and has pushed investors from the traditional domestic versus international divide (or regional divide) towards an integrated global perspective. At the same time, there has been a decrease in the relative preference investors have for domestic equities (“home bias”). Institutional investors across the world have gradually recognized the traditional rationale for home bias has weakened. As asset growth is the main objective of equity allocation, biasing it towards the domestic market may come with

potentially huge opportunity costs. In this new paradigm, the full equity opportunity set spans the globe and is the natural starting point for equity allocation.

(16)

Section II: Alternative Weighting Schemes:

Market-Cap Weighting versus GDP Weighting

Having established in Section I the rationale for using the global opportunity set as the starting point for asset allocation, in this section we examine the rationale for alternative weighting schemes to

determine the geographic asset allocation, with a focus on the use of GDP weights to set the country weights in the equity portfolio.

Market Capitalization Weighting versus Alternative Weighting Schemes

A capitalization-weighted index is an index whose components are weighted according to the total market value of their outstanding shares. The impact of a component's price change is proportional to the issue's overall market value, which is the share price times the number of shares outstanding. There are many reasons why market-cap-weighting has been the dominant standard for stock market indexes. First, market-cap-weighted indices offer an objective way to describe the composition of the

opportunity set using capitalization as the proxy for size. Second, they are relatively simple to calculate since the weightings of index components automatically adjust as stock prices change daily. Market-cap-weighting is also consistent with a passive buy and hold strategy. Third, the Market-cap-weighting scheme favors stocks with higher trading liquidity and capacity since it tilts towards larger companies.

However, in the presence of price inefficiency and mean reversion, the capitalization weighting scheme can be less than desirable when price bubbles and other temporary price disequilibria will affect weights. This may result in performance drag and volatility of security and sector weights resulting in excessive volatility in returns.

Alternative weighting schemes have been proposed which do not use prices in the weighting. An equal-weighting scheme for instance results in less volatile industry weights. Other alternative equal-weighting schemes which aim to reduce volatility include risk-weighting, minimum volatility, and diversity-weighted. Other weighting schemes aim to enhance returns. One example is weighting based on

fundamental characteristics such as price-to-book-value. These and other weighting schemes are further discussed in Melas, Briand, and Urwin (2011).

GDP Weighting of Global Equities

Market-cap-weighted indices reflect the available investment opportunity set in public equity markets. By design, they ignore any unlisted companies, whether privately held or state-owned, since these are not accessible to the investing public. One of the oldest alternative weighting schemes is one that weights countries by their Gross Domestic Product (GDP). Consequently, the weights of countries in the GDP-weighted index will represent the relative importance of a country’s economy as opposed to the size of its equity market.

MSCI GDP Weighted Indices were first launched more than 20 years ago. The MSCI World GDP Index, reflecting developed markets, aimed at the time to address the issue of the large weight of Japan in the MSCI World Index. Later, the GDP-weighted indices were extended to cover emerging markets. Within emerging markets, weighting based on economic size rather than market capitalization was a way to address the divergence between economic size and market size of many countries with the faster

(17)

growing economies. The idea was that developing countries would progressively adopt market-oriented policies in a globalizing world. For further discussion of emerging markets, see Section V.

The GDP-weighting scheme overweights (underweights) countries with economic weight greater

(smaller) than the market capitalization weight. According to the MSCI GDP weighting methodology, the weights of the countries in the index are set in May of each year to each country’s nominal GDP as a percentage of the total. Note that between the May rebalancing, the country weights naturally evolve with changes in the countries’ prices and market capitalizations.

Exhibit 12 summarizes the effects of GDP-weighting. Within developed markets, the largest overweights are to Japan, Germany, and Italy. Meanwhile the largest underweights are to the USA, UK, and

Switzerland. When emerging markets are added to the developed markets universe, the largest overweight becomes China, with India, Russia, Brazil, and Mexico all appearing in the top 10 overweights. In both cases, GDP-weighting implies a greater allocation towards Europe.

Exhibit 12: Top 10 Over-and Underweighted Countries in MSCI GDP-Weighted Indices

Source: MSCI. Data as of November 30, 2011.

Over time, of course, these weights have not remained constant. Exhibit 13 displays the difference between the GDP weight and the market capitalization weight of select countries and emerging markets in MSCI ACWI.

Japan began in 1988 as a large underweight in the GDP index – its GDP weight was 19.3% while its market-capitalization weight is 40.8%. However, after the burst of the asset price bubble, the

Developed Markets Only

MSCI World GDP-Weighted Index MSCI World Index Difference

Developed and Emerging Markets MSCI ACWI GDP-Weighted Index MSCI ACWI Index Difference JAPAN 15.34% 9.36% 5.98% CHINA 9.83% 2.23% 7.60% GERMANY 7.62% 3.46% 4.16% JAPAN 10.68% 8.16% 2.52% ITALY 4.20% 0.99% 3.21% GERMANY 5.31% 3.02% 2.29% SPAIN 3.23% 1.39% 1.84% ITALY 2.92% 0.86% 2.06% FRANCE 5.68% 3.88% 1.80% INDIA 2.68% 0.85% 1.82% NETHERLANDS 1.86% 1.04% 0.81% RUSSIA 2.58% 0.87% 1.71% BELGIUM 1.11% 0.41% 0.71% BRAZIL 3.46% 1.89% 1.57% NORWAY 1.05% 0.40% 0.65% MEXICO 1.86% 0.59% 1.27% AUSTRIA 0.69% 0.10% 0.59% SPAIN 2.25% 1.21% 1.04% IRELAND 0.56% 0.11% 0.45% INDONESIA 1.37% 0.37% 0.99% USA 39.46% 52.07% -12.61% USA 27.49% 45.41% -17.92%

UNITED KINGDOM 5.87% 9.77% -3.90% UNITED KINGDOM 4.09% 8.52% -4.44%

SWITZERLAND 1.28% 3.60% -2.33% SWITZERLAND 0.89% 3.14% -2.25%

CANADA 3.85% 5.16% -1.31% CANADA 2.68% 4.50% -1.82%

HONG KONG 0.56% 1.21% -0.65% AUSTRALIA 2.24% 3.28% -1.05%

AUSTRALIA 3.21% 3.76% -0.55% TAIWAN 0.70% 1.39% -0.69%

SWEDEN 1.00% 1.24% -0.24% HONG KONG 0.39% 1.05% -0.66%

SINGAPORE 0.54% 0.73% -0.19% SWEDEN 0.70% 1.08% -0.38%

FINLAND 0.51% 0.39% 0.12% SOUTH AFRICA 0.65% 0.97% -0.32%

ISRAEL 0.47% 0.27% 0.20% SINGAPORE 0.38% 0.64% -0.26%

Dev. Europe ex UK/Switz. 44.04% 26.54% 17.51% Dev. Europe ex UK/Switz. 21.25% 12.64% 8.61% Top 10 Overweights

(18)

underweight was progressively reduced. Today, Japan has the largest overweight amongst developed markets.

Some countries have had more stable differences between economic and market capitalization weights throughout the history of MSCI ACWI. The large economies of continental Europe, such as Germany, Italy, and France have been persistently over-weighted in the MSCI ACWI GDP Weighted Index. This may be explained by a relatively high proportion of companies not publicly listed in these countries.

Meanwhile, the US has consistently been underweighted throughout history.

Exhibit 13: Difference between GDP and market capitalization weights of select countries and emerging markets in MSCI ACWI (1988-2010)

Source: GDP figures from the World Bank. Market capitalization weights from MSCI. A note on timing: GDP for 2010 is reported in June 2011. It is compared against market cap as of June 2011.

Since 1987, the overweight of emerging market countries in MSCI ACWI has grown from 6.8% to 18.2%. The EM GDP weight has been growing significantly faster than the market capitalization weight.

-25% -20% -15% -10% -5% 0% 5% 10% 15% 20% 25% A ct iv e W e ig h ts ( A C W I G D P W e ig h te d M in u s A C W I M a rk e t C a p W e ig h te d )

CHINA GERMANY SWITZERLAND

USA BRAZIL FRANCE

(19)

Risk and Return

The different weight distribution has led to long-term performance differentials between GDP-weighted indices and their market capitalization weighted counterparts. Exhibit 14 shows the performance of three GDP weighted indices (MSCI World, MSCI EM, and MSCI ACWI GDP Weighted Indices) relative to their market capitalization weighted benchmarks.

Exhibit 14: Relative performance of the MSCI World, MSCI EM, and MSCI ACWI GDP-Weighted Indices

Source: MSCI. Monthly data as of December 31, 2011 in USD. MSCI EM and MSCI ACWI GDP Weighted Indices are simulated before 2005. The MSCI World GDP Weighted Index is simulated before 1988.

Over the history, all three GDP weighted indices have outperformed their market capitalization

weighted counterparts. The effect of the Japanese asset price bubble is particularly striking at the end of the 80s: the GDP-weighted variant of the MSCI World Index underperforms and then sharply

outperforms its market capitalization weighted counterpart. Also, the impact of adding emerging markets is striking.

Exhibit 15 shows performance measures by sub period. The only decade where the GDP-weighting scheme underperforms the market-cap-scheme is in the 1980s. Even then, the return-to-risk ratio remains higher. 80 100 120 140 160 180 200 220 240 260 280

MSCI World GDP / MSCI World MSCI EM GDP / MSCI EM MSCI ACWI GDP / MSCI ACWI

(20)

Exhibit 15: Performance of Global GDP-Weighted and Market Cap-Weighted Indices (January 1970 to December 2011)

The MSCI ACWI GDP-Weighted Index and MSCI ACWI are shown here. Performance is in USD. Prior to December 1987, the MSCI World Index is used in place of the MSCI ACWI. Prior to July 2000, the MSCI World GDP Index is used in place of the MSCI ACWI GDP Index. Geometric average annualized returns are shown. Source: MSCI.

In evaluating alternative weighting schemes, investors may also be interested in characteristics such as downside tail risk, betas, and correlation to the market. As seen in Exhibit 16, a beta and correlation close to 1 suggests that the GDP-weighted index behaves very similarly in these respects to the MSCI ACWI. Moreover, the measures of extreme risk, both empirical conditional Value-at-Risk (VaR)10 and the average returns during the crises periods shown are also quite comparable.

10

Conditional Value-at-Risk is also known as Expected Shortfall. We calculate this empirically as the average monthly loss (not annualized) once the VaR threshold of 95% or 99% has been exceeded.

Market-Cap Weighted Index GDP-Weighted Index

Annualised Return 9.3% 10.5% Annualised Risk 15.3% 15.4% Return-to-Risk 0.61 0.68 1970-1979 Annualised Return 7.0% 8.3% Annualised Risk 14.1% 13.4% Return-to-Risk 0.50 0.62 1980-1989 Annualised Return 20.0% 19.8% Annualised Risk 14.7% 14.4% Return-to-Risk 1.36 1.38 1990-1999 Annualised Return 11.7% 13.2% Annualised Risk 14.1% 13.4% Return-to-Risk 0.83 0.99 2000-December 2011 Annualised Return 1.2% 2.8% Annualised Risk 17.4% 18.9% Return-to-Risk 0.07 0.15

(21)

Exhibit 16: Extreme Risk, Betas, and Correlations

Note: Geometric average monthly return in USD, is shown above. The MSCI ACWI GDP-Weighted Index and MSCI ACWI are shown here. Prior to December 1987, the MSCI World Index is used in place of the MSCI ACWI. Prior to July 2000, the MSCI World GDP Index is used in place of the MSCI ACWI GDP Index. Geometric average monthly return is shown above. Average monthly return is the unannualized geometric avearge of monthly returns. Standard deviation is that of monthly returns, also unannualized. Both skewness and kurtosis are defined in the standard way; kurtosis is excess kurtosis here. Value-at-Risk is the realized monthly percentile loss at the relevant threshold. Conditional Value-at-Risk is the average loss once the threshold has been exceeded. Average negative and positive returns are simple averages conditional on the returns being negative or positive respectively. Source: MSCI

The beta and correlation can of course change over time. In Exhibit 17, we plot rolling three-year trailing beta of the MSCI ACWI GDP-Weighted index relative to the parent ACWI index. We also plot rolling three-year correlations. The beta of the MSCI ACWI GDP has ranged roughly between 0.8 and 1.2, peaking during the 2008 Financial Crisis. The correlation between MSCI ACWI GDP and MSCI ACWI has historically been well above 0.9. Since 1999, it has been above 0.99.

Market-Cap Weighted Index

GDP-Weighted Index

Average Monthly Return 1.0% 1.0%

Standard Deviation 4.2% 4.0% Skewness -0.3 -0.5 Kurtosis 1.6 2.3 VaR (95%) -5.9% -5.4% VaR (99%) -10.5% -10.3% Conditional VaR (95%) -9.0% -9.3% Conditional VaR (99%) -13.5% -13.8%

Average Negative Return -3.0% -2.7%

(22)

Exhibit 17: Historical Betas and Correlations of the GDP-Weighted Index to Market-Cap Weighted Index (Monthly Returns in USD)

The MSCI ACWI GDP-Weighted Index and MSCI ACWI are shown in the exhibit above. Prior to December 1987, the MSCI World Index is used in place of the MSCI ACWI. Prior to July 2000, the MSCI World GDP Index is used in place of the MSCI ACWI GDP Index. Source: MSCI

Attributing Returns

Next we explore which countries contributed the most to the outperformance of the GDP-weighted index over the long run. To do so, we use the Barra Global Equity Model (GEM2) to attribute returns across the various countries. The model defines 134 factors – a World factor, 55 country factors, 8 style factors, 34 industry factors, and 55 currency factors – which capture systematic return globally. These factors are net of each other, estimated in a multivariate cross-sectional regression. The return

contributions thus sum to the overall portfolio return. (Appendix B provides more detailed descriptions of the model.)

Exhibit 18 first looks at the country tilts which contributed the most to the outperformance of the MSCI ACWI GDP-Weighted Index over the MSCI ACWI Index. Since 1988, the largest contributor to return was the overweight to Emerging Markets.

0.60 0.70 0.80 0.90 1.00 1.10 1.20 1.30 Ja n -7 3 Ja n -7 5 Ja n -7 7 Ja n -7 9 Ja n -8 1 Ja n -8 3 Ja n -8 5 Ja n -8 7 Ja n -8 9 Ja n -9 1 Ja n -9 3 Ja n -9 5 Ja n -9 7 Ja n -9 9 Ja n -0 1 Ja n -0 3 Ja n -0 5 Ja n -0 7 Ja n -0 9 Ja n -1 1

Beta of MSCI ACWI GDP to MSCI ACWI

(23)

Exhibit 18: Countries That Drove Performance (January 1997 to December 2011)

* The returns shown are geometric average returns to the MSCI indices, in USD. Source: MSCI

Note that the country contributions to return have not been the same in every time period. For instance, in the 1990s, the US underweight dragged down the performance of the MSCI ACWI GDP-Weighted Index. During the 2000s, the overweight to Europe dragged performance down.

Alongside countries, the impact of relative tilts towards sectors and risk premia or styles can also help or hinder return. Exhibit 19 summarizes the contribution to return from the eight risk premia or style factors from the model as well as from the 10 GICS sectors. As with the country factors, these factors capture the return to stocks in a particular industry or with certain fundamental characteristics, net of the other factors in the model. Thus for instance, the return from the Value factor is the return to the index arising from a tilt towards stocks with high book-to-price, after accounting for the stocks’ industries, country memberships, and other style characteristics. Note that with the exception of the Momentum and Volatility style factors, stock-level exposures to factors are standardized locally within the company’s country of membership. This allows direct comparison of large cap performance in the US versus large cap performance in a smaller emerging market where market cap size is lower on average.

A few factors stand out in Exhibit 19. The MSCI ACWI GDP on average underweighted momentum stocks more than the MSCI ACWI which helped its performance. The GDP-weighted index was also overweight value stocks which also improved returns.

Exhibit 19: The Effect of Risk Premia and Sectors on Returns (January 1997 to December 2011)

Source: MSCI. Performance is in USD.

US Europe Pacific EM

1988 to 2011

Average Weight Difference (%) -13.39% 9.65% -1.05% 12.22%

Annualized Average Returns* 9.56% 10.18% 9.36% 12.60%

January 1988 to December 1999

Average Weight Difference -9.57% 11.96% -1.03% 9.42%

Annualized Average Returns* 19.73% 13.78% 13.42% 17.06%

January 2000 to December 2011

Average Weight Difference -17.89% 6.92% -1.06% 15.52%

Annualized Average Returns* 0.20% 1.60% -0.22% 8.29%

Average Active Exposure Contribution to Return Average Active Exposure Contribution to Return

Style Factors (Risk Premia) Industries

Momentum -0.05 0.18% Energy 0.50 0.05%

Volatility 0.05 -0.02% Materials 0.95 -0.02%

Value 0.04 0.20% Industrials 0.31 -0.02%

Size -0.05 0.02% Consumer Discretionary 0.17 -0.07%

Size Nonlinearity 0.03 0.03% Consumer Staples -1.22 -0.03%

Growth 0.00 -0.02% Health Care -2.70 -0.08%

Liquidity 0.04 0.01% Financials 0.87 -0.08%

Financial Leverage 0.03 -0.05% Information Technology -1.84 -0.16%

Telecommunication Services 2.07 -0.06%

(24)

However in general, the return impact of risk premia and sectors is considerably lower than the country effects discussed earlier. The relative importance of these different sources of return is summarized in Exhibit 20. Overall, GDP weighting seems to be an active bet on the country factors with much less impact coming from risk premia or industry tilts. In other words, the outperformance of the GDP-weighted index has historically not come from overweighting small caps or high book-to-price stocks (defined as small caps or high book-to-price within their own markets) so much as it has arisen from differences between the countries themselves.

Exhibit 20: Country Overweights and Underweights Are the Main Drivers of Systematic Returns (January 1997 to December 2011)

Source: MSCI

The results discussed in Exhibits 19 and 20 are of course dependent on the modeling assumptions. In particular, the fact that Size and Value characteristics are standardized within countries raises the question of whether GDP weighting results in under- or over-weighting countries that are themselves on average smaller or have better valuations. In Exhibit 21, we show the countries with the main

countries/regions along with their average price-to-book-value and market capitalization over various periods. We find that countries which are over-weighted in the GDP scheme tend to have lower price-to-book-value and smaller market cap compared to those that are under-weighted. Thus, it may be that the country factors in part have captured the differences in valuation and size. Alternatively, the return not attributed to any of the factors (i.e., Specific Return in Exhibit 20) may have partly captured these differences.

Exhibit 21: GDP Weights, Price-to-Book Value, and Market Capitalization (Monthly Averages, January 1997 to December 2011)

Source of Return

Return Contribution

Portfolio Return (MSCI ACWI GDP) 6.68%

Benchmark (MSCI ACWI) 4.07%

Active Return 2.61%

Country Factors 1.19%

Style Factors (Risk Premia) 0.35%

Industry Factors -0.49% Specific 1.41% Average Active Weight Average P/B Average Active Weight Average P/B Average Active Weight Average P/B Average Active Weight Average P/B Average Active Weight Average P/B USA -13.4% 2.88 -6.2% 2.25 -17.1% 4.42 -19.8% 3.03 -16.2% 2.44 UK -4.6% 2.31 -4.4% 2.01 -4.6% 3.31 -5.3% 2.15 -4.1% 2.11 JAPAN -2.5% 2.22 -11.0% 3.14 4.1% 2.05 3.4% 1.65 0.4% 1.54

DEV. PACIFIC ex JAPAN -1.0% 1.89 -1.3% 1.76 -0.4% 1.93 -0.6% 1.87 -1.7% 2.13

CANADA -0.2% 2.00 0.4% 1.60 0.0% 2.28 -0.2% 2.08 -1.5% 2.35

DEV. EUROPE ex UK 9.7% 2.11 15.0% 1.66 4.7% 3.12 7.4% 2.25 7.4% 1.90

EMERGING MARKETS 12.2% 1.89 7.8% 1.84 13.2% 1.69 15.3% 1.80 16.0% 2.22

1987-1995 1996-2000 2001-2005 2006-2010

(25)

Source: MSCI

A Discussion of GDP Weighting

A variety of reasons have been proposed as to why GDP weighting has historically resulted in higher returns. Proponents of GDP weighting first argue that GDP weights serve as a better proxy of the natural country weight in the CAPM total market portfolio. According to the CAPM, investors should own the market portfolio which includes all holdings by investors, listed and unlisted. Holding countries in proportion of GDP may bring the weights of countries in the portfolio closer to the natural country weight as it incorporates the non-listed portion of the economy.11

Second, GDP weighting is not dependent on stock prices which means countries which experience a temporary bubble will not be as heavily weighted as they would be in a market-cap-weighting scheme. Recall that the MSCI World GDP-Weighted Index was originally created in response to the significant asset bubble in Japan in the 1980s.

Third, GDP weighting has historically provided increased exposure to emerging markets thus benefiting from a significant emerging market risk premium. Advocates of this line of reasoning argue that as these markets progressively open up, their equities will attract inflows that may result in a virtuous cycle. As a result, the market capitalization to GDP ratio of these countries would increase, in part due to above average returns. This implies that GDP weighting could imply a tactical tilt towards these countries in expectation of higher returns.

For detractors of GDP-weighting, several counterarguments exist. First, it is not true that all countries with room to grow their market size end up realizing their potential. Not all countries with relatively larger economies compared to the size of their markets successfully experience strong market growth. Moreover, not all emerging markets for instance have had an easy time liberalizing, take for instance the case of Argentina. As countries seek to expand, political forces and unexpected shifts in consumer demand and production supply have been known to disrupt the process.

Overall, there does appear to be a relationship between GDP growth and equity returns. Dimson, Elroy, and Marsh (2010) for instance find a positive correlation of 0.41 between real GDP growth rates and real equity returns (for developed markets).12 However, it is important to note that historically the

relationship has not been one-for-one. As Arnott and Bernstein (2003) pointed out, the growth of listed

11_{It is important to note that the MSCI indices uses free float market capitalization in constructing market capitalization based weights. This raises the question of} how similar GDP-weighted indices are to indices constructed using the full market capitalization for each country. A detailed discussion appears in Appendix C. In all, the correlation between full market cap-derived weights and GDP-weights is not one-for-one. Note however that the free float adjustment impacts the emerging markets much more than developed markets, implying larger weights in emerging markets if a full market capitalization scheme is used. This is qualitatively similar to the results of GDP weighting.

12_{On the other hand, they find a negative correlation between real GDP per capita growth and real equity returns.} Average Active Weight Average Market Capitalization (USD Bill.) Average Active Weight Average Market Capitalization (USD Bill.) Average Active Weight Average Market Capitalization (USD Bill.) Average Active Weight Average Market Capitalization (USD Bill.) Average Active Weight Average Market Capitalization (USD Bill.) USA -13.4% 6,765.2 -6.2% 2,133.2 -17.1% 7,102.4 -19.8% 9,157.9 -16.2% 11,295.2 UK -4.6% 1,443.8 -4.4% 571.0 -4.6% 1,451.1 -5.3% 1,823.3 -4.1% 2,444.7 JAPAN -2.5% 1,956.5 -11.0% 1,821.7 4.1% 1,970.4 3.4% 1,627.5 0.4% 2,452.9

DEV. PACIFIC ex JAPAN -1.0% 582.2 -1.3% 254.9 -0.4% 518.9 -0.6% 541.5 -1.7% 1,105.6

CANADA -0.2% 458.6 0.4% 146.7 0.0% 334.6 -0.2% 445.3 -1.5% 989.6

DEV. EUROPE ex UK 9.7% 2,826.6 15.0% 964.7 4.7% 3,221.5 7.4% 3,295.8 7.4% 4,945.8

EMERGING MARKETS 12.2% 1,172.6 7.8% 259.3 13.2% 876.2 15.3% 883.8 16.0% 2,714.2

(26)

companies contributes only a part of a nation’s increase in GDP. Strong GDP growth may translate into new enterprises, the raising of additional capital, or state or private shareholders selling their stakes, not necessarily market returns.

A final point is that additional consideration should be paid to instances where a country is in the early phases of liberalization. In these cases, implementing the relatively larger weight implied by GDP may run into foreign investability constraints and/or being limited to a narrow slice of the economy. Early investors in China for instance were limited to a small set of stocks. In such cases, investors may bear a large amount of concentration risk as well as stock-specific risk. These cases highlight the point that the GDP-weighting scheme assumes the market is sufficiently representative of the larger economy.

Summary

One of the oldest alternative weighting schemes is one that allocates countries by their Gross Domestic Product (GDP). Within developed markets, the largest overweights are to Japan, Germany, and Italy. Meanwhile the largest underweights are to the USA, UK, and Switzerland. When emerging markets are added to the developed markets universe, the largest overweight becomes China, with India, Russia, Brazil, and Mexico all appearing in the top 10 overweights. Overall, GDP weighting seems to be an active bet on country factors, allocating more to emerging markets and less to developed markets.

Proponents of GDP-weighting highlight several advantages. First, it weights countries based on their economic size, not on the size of their markets, which can sometimes reflect a narrower view of the economy particularly in emerging markets. Second, GDP weighting is not dependent on stock prices which means countries which experience a temporary bubble will not be as heavily weighted as they would be in a market-cap-weighting scheme. Third, GDP weighting has historically provided increased exposure to emerging markets thus benefiting from a significant emerging market risk premium. The main risk however to GDP-weighting is if growth comes not in the form of market growth but is confined to the closed part of the economy.

(27)

Section III: An Examination of Concentration Risk with

a Focus on the US and Europe

Different weighting schemes result in various levels of geographical concentration risk. In this section, we discuss concentration risk with a focus on US and European markets.

The Issue of Geographical Concentration Risk

Market capitalization weighted indices serve as efficient tools to capture the broad equity market beta through an objective representation of the entire opportunity set, macro consistency, automatic

rebalancing, low transaction costs, high trading liquidity, and maximum investment capacity. Per design, market cap indices reflect the size of its constituent companies and/or countries. One potential

disadvantage of market-cap-weighting is that in some cases, a single or small group of stocks can represent a large share of the index. For example, a security that has experienced a large run-up in price will have a much larger weight in a market-cap-weighted index. In such cases, changes in this security’s price will have a necessarily large impact on the index’s return and risk profile. Indices with a large number of names tend to dampen this issue. Hence, the most widely used institutional benchmarks for the US for instance contain upwards of 1,000 names.

At the global level, a similar situation can occur for various countries or regions. A geographically concentrated index will wax and wane with the largest country or countries in the index. The extent to which an index is subject to this dominance can be thought of as geographic concentration risk.

Exhibit 22 shows the ratio of market cap of the MSCI USA Index and the MSCI Europe Index to the global opportunity set (MSCI ACWI). US equity market cap has historically been larger than Europe’s.

Currently, the US represents about 45% of MSCI ACWI raising the question of whether this represents a great deal of geographical concentration risk.

Exhibit 22: MSCI USA and MSCI Europe as a Share of ACWI (Market Capitalization, USD)

Source: MSCI 0% 10% 20% 30% 40% 50% 60%

Dec-87 Dec-89 Dec-91 Dec-93 Dec-95 Dec-97 Dec-99 Dec-01 Dec-03 Dec-05 Dec-07 Dec-09 MSCI USA MSCI Europe

(28)

The notion of concentration risk in the US or Europe is hard to define and encompasses many different aspects. First, we can look at historical volatility and performance of those markets during past crises. We can further examine the drivers of risk and simulate the impact on performance by decreasing and increasing the weight to US and Europe historically. Second, we can also look at the profile of US and European equities today. To what extent are these markets fairly broad and diversified and/or depend on foreign sources of return?

Note in this section we treat Europe as one unit given its integrated economy. While there remains a risk to the continuation of the Euro bloc, at least in its current form, for this analysis we assume that the countries that make up Europe will continue to have close economic ties.

Return and Risk of US and European Equities

Exhibit 23 summarizes the return and risk of US and European equities over the last forty years. Returns in both USD and NOK are shown. Europe posted somewhat higher returns (in both currencies) over this period. In USD, the volatility of European equities was higher and the risk-adjusted return lower. In NOK, the volatility of European equities was lower and the risk-adjusted return higher. Performance varied from decade to decade with Europe experiencing the largest gains relative to the US in the 1980s and 2000s (up until recently).

Exhibit 23: Summary of Historical US and Europe Performance (Monthly Returns in USD and NOK, January 1970 to December 2011)

Note: Geometric average is used for average annualized returns above. Source: MSCI

MSCI USA (USD) MSCI Europe (USD) MSCI USA (NOK) MSCI Europe (NOK)

Average Annualized Returns

1970 - 2011 9.5% 10.1% 9.0% 9.6% 1970-1979 4.6% 8.6% 0.8% 4.6% 1980-1989 17.1% 18.5% 20.6% 22.0% 1990-1999 19.0% 14.5% 21.4% 16.8% 2000-2009 -1.3% 2.4% -4.5% -0.9% 2010-2011 8.5% -3.3% 10.3% -1.7%

Annualized Standard Deviation

1970 - 2011 15.7% 17.8% 17.5% 15.7% 1970-1979 15.9% 17.0% 17.2% 15.4% 1980-1989 16.2% 18.0% 19.1% 14.7% 1990-1999 13.4% 14.6% 16.9% 15.4% 2000-2009 16.2% 19.4% 17.0% 17.0% 2010-2011 17.5% 24.6% 11.2% 13.3%

Risk-Adjusted Annualized Return

1970 - 2011 0.60 0.57 0.51 0.61 1970-1979 0.29 0.51 0.05 0.30 1980-1989 1.05 1.02 1.08 1.49 1990-1999 1.42 0.99 1.27 1.09 2000-2009 -0.08 0.13 -0.26 -0.05 2010-2011 0.49 -0.13 0.92 -0.13

(29)

A few measures of downside risk and tail risk are shown in Exhibit 24. Along all measures, European equities have historically had slightly greater downside risk and tail risk. The differences between Europe and US however are relatively small.

Exhibit 24: Downside and Tail Risk Measures (Monthly Returns in USD, January 1970 to December 2011)

Note: Geometric average monthly return is shown above. Average monthly return is the geometric average of monthly returns not annualized. Standard deviation is that of monthly returns, also not annualized. Both skewness and kurtosis are defined in the standard way; kurtosis is excess kurtosis here. Value-at-Risk is the realized monthly percentile loss at the relevant threshold. Conditional Value-at-Risk is the average loss once the threshold has been exceeded. Average negative and positive returns are simple averages conditional on the returns being negative or positive respectively. Source: MSCI

If we look next at performance in major crises, again the differences between the US and Europe are relatively minor. As shown in Exhibit 25, the Japan crash, Dot com crash and Subprime crisis had effects across markets.13 In contrast, the effects of the Sovereign Debt Crisis in Europe have been

unprecedented historically over the period and have affected European equity markets much more so than other regions.

Exhibit 25: Performance in Past Periods of Market Stress (Cumulative Return During Period, in Percentages)

A. USD

13_{The Asian/LTCM period shows positive returns for the MSCI USA and MSCI Europe because the period encompasses the rebound.} MSCI USA (USD) MSCI Europe (USD) MSCI USA (NOK) MSCI Europe (NOK)

Average Monthly Return 0.8% 0.8% 0.7% 0.8%

Standard Deviation 4.5% 5.1% 5.1% 4.5% Skewness -0.4 -0.4 -0.2 -0.5 Kurtosis 1.8 1.9 1.5 1.8 VaR (95%) -7.2% -8.3% -7.1% -7.2% VaR (99%) -10.5% -13.1% -11.4% -11.2% Conditional VaR (95%) -9.7% -10.5% -10.3% -10.1% Conditional VaR (99%) -14.3% -16.3% -15.1% -14.1%

Average Negative Return -3.4% -4.0% -3.9% -3.6%

Average Positive Return 3.6% 3.9% 4.0% 3.6%

Market Stress Episodes

1988 to 2011 MSCI USA MSCI Europe MSCI Japan MSCI EM

Japan Crash -10.3 -10.6 -44.1 -7.7

Asian Crisis/ LTCM 25.3 28.0 -4.5 -22.4

Dot Com Crash -43.3 -45.1 -58.2 -41.5

Subprime Crisis -50.6 -59.0 -45.4 -61.4

(30)

B. NOK

The periods are defined as follows: (1) Japan Crash: 12/1989 to 9/1990, (2) Asian Crisis: 11/1997 to 11/1998, (3) Dot Com Crash: 3/2000 to 3/2003, (4) Subprime Crisis: 10/2007 to 2/2009, (5) Europe Sovereign Crisis 1/2010 to 11/2011. Source: MSCI

We next evaluated performance across different types of macroeconomic and market regimes: (1) recession; (2) market volatility (high and low); and (3) asset price bubbles (high and low). Appendix D contains detailed results of this analysis. To highlight some of our findings, we find that historical performance in recession periods has been relatively close; for Europe, average (simple) returns using USD monthly data were 0.01 bps during recessions compared to -0.21 bps for the US. In high volatility regimes (when the level of the VIX is rapidly rising), we find average returns to the US and Europe of -1.9% and -2.4% respectively. Moreover, the historical spread in returns between high and low volatility regimes is relatively close for the two markets.

Simulating Over- and Under-weights to the US and Europe

An intuitive way to study the impact of changing the allocation to US and Europe is to simulate historical performance under different weighting schemes. For instance, we can directly test the results had the global equity portfolio historically pulled back the significant market cap weight to the US. Here we consider two cases, the first case being the simpler of the two approaches.

• Case 1: Each month, we take the North American market cap weight and subtract 10 percentage points and add 10 percentage points to Europe

• Case 2: Each month, we apply a factor of 1.5 to the Europe market cap weight and a factor of 1.0 to North America market cap weight. The new weight in each month is:

0 . 1 * 5 . 1 * * 5 . 1 * , , , , , , old NorthAm old Europe old NorthAm old Europe old Europe new Europe w w w w w w + + = (3.1) 0 . 1 * 5 . 1 * * 0 . 1 * , , , , , , old NorthAm old Europe old NorthAm old Europe old NorthAm new NorthAm w w w w w w + + = (3.2)

In both cases we fix the weight of the Pacific region (developed Pacific and Asia countries) as well as all emerging markets. Exhibit 26 shows the results of the simulations and compares them to the market cap weighted global index (MSCI ACWI) as well as the MSCI ACWI GDP-weighted index from Section II. First, we note that Case 1 and Case 2 deliver very similar returns. Second, we see that both simulations outperform MSCI ACWI historically but neither outperform the MSCI ACWI GDP-weighted index.14

14

Recall that from Exhibit 2, European equities outperformed US equities from 2000 to 2009 but underperformed during the 1990s and 2010-2011 so it is unclear a priori what the effect of giving a higher weight to European equities would have been.

Market Stress Episodes

1988 to 2011 MSCI USA MSCI Europe MSCI Japan MSCI EM

Japan Crash -17.7 -17.9 -48.7 -15.3

Asian Crisis/ LTCM 30.0 32.9 -0.9 -19.5

Dot Com Crash -51.2 -52.8 -64.1 -49.8

Subprime Crisis -35.5 -46.4 -28.6 -49.6

(31)

Exhibit 26: Simulation Results (Global Portfolios with Varying North American and European Weights, January 1988 to December 2011, Monthly Returns)

A. USD

B. NOK

Source: MSCI

Market Cap Weight

(ACWI) GDP Weight (ACWI)

Simulation Results (Case 1)

Simulation Results (Case 2) Average Annualized Return

1988-2011 7.0% 8.7% 7.7% 7.7%

- January 1988 to December 2000 10.8% 12.0% 11.6% 11.7%

Average Standard Deviation

1988-2011 15.7% 16.3% 16.0% 16.0%

1988-2011 0.45 0.53 0.48 0.49

- January 1988 to December 2000 0.77 0.89 0.82 0.83

Average Weight to North America 45.4% 32.1% 35.4% 38.2%

Average Weight to Europe 27.5% 32.4% 37.5% 34.7%

Average Weight to Pacific 20.9% 17.3% 20.9% 20.9%

Average Weight to Emerging Markets 6.2% 18.2% 6.2% 6.2%

Market Cap Weight

(ACWI) GDP Weight (ACWI)

Simulation Results (Case 1)

Simulation Results (Case 2) Average Annualized Return

1988-2011 6.8% 8.3% 7.5% 7.5%

- January 2001 to December 2011 -1.2% 0.9% -0.7% -0.7%

Average Standard Deviation

1988-2011 15.5% 15.6% 15.4% 15.4%

1988-2011 0.44 0.53 0.48 0.49

- January 2001 to December 2011 -0.08 0.05 -0.05 -0.05

Average Weight to North America 45.4% 32.1% 35.4% 38.2%

Average Weight to Europe 27.5% 32.4% 37.5% 34.7%

Average Weight to Pacific 20.9% 17.3% 20.9% 20.9%

(32)

Attributing Return and Risk

Overall return and risk characteristics between the US and Europe have been historically similar, but are there differences in the drivers of returns? Next, we use the Barra Global Equity Model (GEM2) to attribute historical returns and risk. There are four main categories for attribution: Currencies,

Countries, Styles (or risk premia) and Industries. A separate World factor reflects a “market” factor. The model estimates factors that reflect the pure return to these categories and can be used to attribute the performance of any index or portfolio. All factors represent systematic sources of return (akin to beta) and are non-diversifiable. (Appendix B describes the model in detail.) Exhibit 27 shows the return and risk from each category over the period January 1997 to December 2011. The single largest contributor to return and risk has been the World factor. The risk from the World factor – 17% and 15.4% for Europe and US, respectively—makes up the majority of risk of the two entities. In other words, the bulk of risk was completely non-diversifiable even if country weights were varied; the US and Europe share a significant joint source of systematic risk.

Exhibit 27: Attribution of Return for US and Europe (January 1997 to December 2011, Barra Global Equity Model, Local Returns)

Source: MSCI

After the World factor, it is Country factors that next dominate risk (at 5.9% and 5.4% annualized contributions respectively for Europe and the US).15 Country risk for Europe is an aggregation of the 20 country factors representing the markets that make up the European market. By far the most volatile European country factors over this period were the UK country factor followed by France and Germany. The trailing 12-month standard deviation of these factors is plotted alongside that of the US country factor in Exhibit 28.

15

Country risk was highest over this time period but note that the return from style and industry factors was not far off from country factors. This conforms to our discussion in Section I in which we pointed out that country membership no lo